| 00:01.06 | *** join/#storm wallflower (n=wallflow@ip-205-246-113-216.pool.grokthis.net) |
| 05:14.38 | *** join/#storm jukart (i=lovely@lsfw01.lovelysystems.com) |
| 05:22.54 | *** join/#storm jukart_ (i=lovely@lsfw01.lovelysystems.com) |
| 06:07.46 | *** join/#storm jukart (i=lovely@lsfw01.lovelysystems.com) |
| 06:28.31 | *** join/#storm dobee (i=lovely@lsfw01.lovelysystems.com) |
| 09:56.59 | *** join/#storm thekorn (n=mkorn@studserv.stud.uni-hannover.de) |
| 09:59.04 | thekorn | hi, how do I store datetime objects with timezone information in a sqlite db? |
| 09:59.51 | thekorn | when I try to read such value I get a ValueError: |
| 09:59.55 | thekorn | Unknown time format: u'17:41:27.985926+00:00' |
| 10:03.21 | therve | I think youneed to use a TEXT field |
| 10:07.29 | thekorn | hmm, both VARCHAR and TEXT don't work, same error |
| 10:12.28 | therve | do you have an example script? |
| 10:34.38 | thekorn | therve: yes, let me write a minimal example |
| 10:44.41 | thekorn | therve: http://paste.ubuntu.com/55566/ |
| 10:46.33 | therve | thekorn: cool, I'm looking at it |
| 10:48.21 | thekorn | thanks |
| 10:56.21 | therve | oh |
| 10:56.29 | therve | # TODO Add support for timezones. |
| 10:56.34 | therve | that must be the problem... |
| 10:59.00 | thekorn | therve: hmm, ok, thanks for investigating |
| 10:59.21 | thekorn | is there a open bug for this? |
| 10:59.33 | thekorn | or something similar, so I can track progress |
| 11:00.53 | therve | let me look |
| 11:02.42 | therve | thekorn: doesn't look like it |
| 11:03.02 | therve | if you can open a bug, that would be great |
| 11:03.20 | thekorn | ok, thanks will do after lunch |
| 11:03.32 | therve | thanks |
| 11:40.32 | *** join/#storm andrea-bs (n=andrea@ubuntu/member/beeseek.developer.andrea-bs) |
| 11:54.31 | *** join/#storm niemeyer (n=niemeyer@200-181-177-142.ctame705.dsl.brasiltelecom.net.br) |
| 11:54.40 | *** join/#storm niemeyer_ (n=niemeyer@200-181-177-142.ctame705.dsl.brasiltelecom.net.br) |
| 11:57.51 | *** join/#storm dobee (i=lovely@lsfw01.lovelysystems.com) |
| 13:54.42 | *** join/#storm Sidnei (n=Sidnei@200-203-77-184.cslce701.dsl.brasiltelecom.net.br) |
| 14:52.06 | *** join/#storm kov (n=kov@debian/developer/kov) |
| 16:04.54 | *** join/#storm andrea-bs (n=andrea@ubuntu/member/beeseek.developer.andrea-bs) |
| 17:02.53 | *** join/#storm dobee (n=dobee@85-124-200-100.static.xdsl-line.inode.at) |
| 17:04.34 | *** join/#storm jdo (n=jdobrien@238.205.119.70.cfl.res.rr.com) |
| 17:09.56 | *** join/#storm jdo (n=jdobrien@238.205.119.70.cfl.res.rr.com) |
| 17:38.24 | *** join/#storm jamesh (n=james@canonical/launchpad/jamesh) |
| 18:20.59 | *** join/#storm djfroofy (n=drew@157.166.167.129) |
| 18:36.38 | *** join/#storm dobee (n=dobee@85-124-200-100.static.xdsl-line.inode.at) |
| 19:41.43 | *** join/#storm dobee (n=dobee@85-124-200-100.static.xdsl-line.inode.at) |
| 19:44.09 | *** join/#storm jdo (n=jdobrien@238.205.119.70.cfl.res.rr.com) |
| 20:20.43 | *** join/#storm jamesh_ (n=james@canonical/launchpad/jamesh) |
| 20:33.23 | *** join/#storm jdo (n=jdobrien@238.205.119.70.cfl.res.rr.com) |
| 20:51.46 | *** join/#storm jamesh__ (n=james@canonical/launchpad/jamesh) |
| 20:59.03 | djfroofy | if a store is rolled back, should that not invalidate its cache? |
| 21:02.28 | niemeyer | djfroofy: It should |
| 21:02.39 | niemeyer | djfroofy: and it does |
| 21:03.05 | niemeyer | djfroofy: I mean, it *should* and it *does* invalidate its cache |
| 21:03.26 | niemeyer | djfroofy: The question was using the negative, so I'm not sure about which way it goes |
| 21:05.19 | djfroofy | niemeyer: just to be clear, if there were an object of type Thingy stored in the database ... |
| 21:05.50 | djfroofy | niemeyer: after rolling back the store, the next query for Thingy would ignore objects cached and use values from database? |
| 21:06.21 | niemeyer | djfroofy: Hold on, there are two different topics playing here |
| 21:06.41 | niemeyer | djfroofy: Both the cached object *and* the values from the database are used |
| 21:07.35 | djfroofy | niemeyer: which means properties are reset on the object using values from the database? |
| 21:07.45 | niemeyer | djfroofy: So, in practice: |
| 21:08.15 | niemeyer | obj.attr = 10; store.commit(); obj.attr = 20; store.rollback(); assert obj.attr == 10 |
| 21:08.31 | niemeyer | djfroofy: Yes, something resembling that |
| 21:09.02 | niemeyer | djfroofy: They're not fetched immediately because they may not be necessary anymore, but they are invalidated, and reloaded when necessary |
| 21:09.45 | djfroofy | niemeyer: well here's the "problem" i'm experiencing, in simple steps ... |
| 21:11.17 | djfroofy | thing = store.find(Thing, Thing.status==2).order_by(...).first(); ... thing.status = 1; store.commit() |
| 21:11.29 | djfroofy | # external process updates thing.status to 1 |
| 21:12.01 | djfroofy | # woops, correction: external process updated thing.status to 2 |
| 21:12.27 | djfroofy | thing = store.find(Thing, Thing.status==2).order_by(...).first(); assert thing.status == 2 |
| 21:12.39 | djfroofy | niemeyer: the assertion failes |
| 21:13.59 | niemeyer | djfroofy: Note that the behavior is transactional |
| 21:14.23 | niemeyer | djfroofy: If you start a transaction before the external process updates it to two, the value is still one in the current transaction |
| 21:16.04 | djfroofy | niemeyer: where are the transaction boundaries? from initialization of Store to descruction of store? or between commits? |
| 21:16.46 | niemeyer | djfroofy: The transaction boundary you can trust is the commit/rolblack |
| 21:16.48 | niemeyer | rollback |
| 21:21.12 | djfroofy | niemeyer: hmmmm ... i'll try putting an isolated test together. because the external update is definitely ocurring before the next transaction |
| 21:21.28 | niemeyer | djfroofy: Ok, that'd help |
| 21:21.38 | niemeyer | djfroofy: Here is a local test to get you started: http://paste.ubuntu.com/55787/ |
| 21:22.51 | djfroofy | niemeyer: thanks |
| 21:52.02 | djfroofy | niemeyer: thanks, my stripped down test just confirms what you've stated about transactions |
| 21:52.50 | djfroofy | i must have a bug in my code somewhere, or there's a bug in twisted-integration |
| 21:56.36 | *** join/#storm jamesh_ (n=james@canonical/launchpad/jamesh) |